Two links from Mark Thoma. First, Clive Crook.

For decades the educational quality of the US labour force surged. In 1940, less than 5 per cent of the population aged 25-64 had at least a four-year college education. By 2000, the proportion had increased to nearly 30 per cent. Successive generations of workers improved on the educational attainments of their predecessors. Retiring workers were replaced by better-educated youngsters. This remorseless accumulation of human capital helped fuel the country’s postwar growth. According to at least one authoritative study, it was the principal driver.

Which authoritative study is that? Is it Goldin-Katz again? They only attribute 14 percent of the productivity increase to more education, and even that amount depends on assuming that all of the wage premium for more education reflects actual schooling rather than ability. But even if you assume the full 14 percent, it is not the “principal driver,” although you could be forgiven for taking that away as from their rhetoric, as opposed to their numbers.

Next, Elizabeth Cascio, Damon Clark, and Nora Gordon,

On average, native-born U.S. teenagers performed worse on the IALS test than teenagers in any other country: Only 4.7% achieved at least level-four proficiency. In contrast, in the highest ranking country, Sweden, more than 35% of students achieved at least level-four proficiency. Across all nations other than the United States, on average 16% of 16- and 17-year-olds were highly skilled…While substantial learning takes place after high school in nearly all countries, these gains are particularly large in the United States. In fact, except for Norway, the age profile of skill is steeper in the United States than in any other country in the sample.

I can’t believe that the authors don’t suggest that U.S. students learn faster out of high school because they have so much more ground to make up. It seems to me like that’s the obvious explanation.